文件名称:Semi-supervised-learning
- 所属分类：
- 人工智能/神经网络/遗传算法
- 资源属性：
- [PDF]
- 上传时间：
- 2012-11-26
- 文件大小：
- 414kb
- 下载次数：
- 0次
- 提 供 者：
- x*
- 相关连接：
- 无
- 下载说明：
- 别用迅雷下载，失败请重下，重下不扣分！
介绍说明－－下载内容均来自于网络，请自行研究使用
义了一个欧氏距离和监督信息相混合的新的最近邻计算函数，从而将K一均值算法很好地应用于半
监督聚类问题。针对K一均值算法初始质心敏感的缺陷，用粒子群算法的搜索空间模拟聚类的欧氏空间，迭代搜
索找到较优的聚类质心，同时提出动态管理种群的策略以提高粒子群算法搜索效率。算法在UCI的多个数据集
上测试都得到了较好的聚类准确率。-Righteousness of a Euclidean distance and supervision of a mixture of new nearest neighbor calculation functions, thus the K-means algorithm applied to the semi-supervised clustering problem. K-means algorithm the initial center of mass-sensitive defect clustering in the search space of the particle swarm algorithm simulation in Euclidean space, an iterative search to find the optimum cluster centroid, and strategies to improve particle swarm optimization to dynamic management of stocks search efficiency. Algorithm on multiple datasets in the UCI tests are a good clustering accuracy.
监督聚类问题。针对K一均值算法初始质心敏感的缺陷，用粒子群算法的搜索空间模拟聚类的欧氏空间，迭代搜
索找到较优的聚类质心，同时提出动态管理种群的策略以提高粒子群算法搜索效率。算法在UCI的多个数据集
上测试都得到了较好的聚类准确率。-Righteousness of a Euclidean distance and supervision of a mixture of new nearest neighbor calculation functions, thus the K-means algorithm applied to the semi-supervised clustering problem. K-means algorithm the initial center of mass-sensitive defect clustering in the search space of the particle swarm algorithm simulation in Euclidean space, an iterative search to find the optimum cluster centroid, and strategies to improve particle swarm optimization to dynamic management of stocks search efficiency. Algorithm on multiple datasets in the UCI tests are a good clustering accuracy.
(系统自动生成,下载前可以参看下载内容)
下载文件列表
基于半监督学习的K-均值聚类算法研究.pdf